Date of Award

Summer 2014

Document Type


Degree Name

Master of Science (MS)


Mechanical Engineering

First Advisor

Schimmels, Joseph M.

Second Advisor

Voglewede,Philip A.

Third Advisor

Rice, James


This thesis presents the design and simulation results of the CamWalk, a novel passive prosthetic ankle that has mechanical behavior similar to that for a natural ankle. The CamWalk uses a compression spring network that allows coupling between two degrees of freedom; one for translation along the leg and another for rotation about the ankle joint. When walking, potential energy from the person's weight is stored in the spring network in deflection along the leg. The energy is released by the network as rotation of the foot. The amount of translational work that is converted to rotational work about the ankle is proportional to the maximum allowed leg deflection, which was limited to 15 mm. A quasi-static model is used to assess the performance of the design and is used in the optimization of the design parameters. Optimizing the design parameters to match the natural ankle characteristics of published average kinetic and kinematic data from gait analyses, yields a design that provides 44.47% of the net rotational work of a natural ankle. Conventional compression springs, used for the spring network of the CamWalk, are interchangeable. These springs are optimized for the individual user, keeping the same prosthesis geometry determined by the optimization for the average walker. Simulation results for three individuals show that spring optimization is sufficient to produce 44.4% (or more) of the natural ankle work. The individual subject results also show that the CamWalk preforms reliably even with variation in the dynamics on the walker. A proof-of-concept prototype was fabricated and tested to verify the quasi-static model accuracy and validate the overall approach. The prototype was "walked" using an industrial robotic manipulator as a positioning source. The deflection and load profiles were measured using potentiometers and a 6-axis force/torque sensor. The prototype's measured rotational work was 93.7% of the work predicted by the quasi-static model, verifying the model's accuracy and demonstrating that energy generated in the deflection is converted into torque about the ankle.